船电技术
船電技術
선전기술
Marine Electric & Electronic Technology
2015年
11期
28-30
,共3页
灰色系统%神经网络%建模%辨识%非线性%估值
灰色繫統%神經網絡%建模%辨識%非線性%估值
회색계통%신경망락%건모%변식%비선성%고치
grey system%neural network%modeling%identification%nonlinear%valuation
采用 BP 神经网络对模型参数进行预测,算法的学习训练速度和建模时间比较长;采用灰色系统理论对模型参数进行预测,对数据信息的学习和训练能力比较有限,两种算法都存在各自的缺陷,为了提高模型中参数的收敛速度和估计精度,本文将灰色系统理论和BP神经网络算法相融合,通过仿真可以看出,模型参数的估值精度比较高,误差较小,证明了该算法的有效性和鲁棒性。
採用 BP 神經網絡對模型參數進行預測,算法的學習訓練速度和建模時間比較長;採用灰色繫統理論對模型參數進行預測,對數據信息的學習和訓練能力比較有限,兩種算法都存在各自的缺陷,為瞭提高模型中參數的收斂速度和估計精度,本文將灰色繫統理論和BP神經網絡算法相融閤,通過倣真可以看齣,模型參數的估值精度比較高,誤差較小,證明瞭該算法的有效性和魯棒性。
채용 BP 신경망락대모형삼수진행예측,산법적학습훈련속도화건모시간비교장;채용회색계통이론대모형삼수진행예측,대수거신식적학습화훈련능력비교유한,량충산법도존재각자적결함,위료제고모형중삼수적수렴속도화고계정도,본문장회색계통이론화BP신경망락산법상융합,통과방진가이간출,모형삼수적고치정도비교고,오차교소,증명료해산법적유효성화로봉성。
When the BP neural network is used to predict the model parameters, the learning training speed and modeling time of algorithm are longer.when the gray system’s theory is used to forecast model parameters, the study and training ability of data information are limited. These two algorithms have their own defects. The grey system’s theory is combined with BP neural network algorithm in this paper to improve the convergence speed and estimation precision in the model. It can be seen from the simulation that the valuation accuracy of model parameter is higher, smaller error, proving the effectiveness and robustness of the algorithm.